Methods

The bucketers of skorecard come with a handy list of methods for you to peek under the hood of the bucketer

from skorecard.datasets import load_uci_credit_card
from skorecard.bucketers import DecisionTreeBucketer

X, y = load_uci_credit_card(return_X_y=True)

specials = {"LIMIT_BAL": {"=50000": [50000], "in [20000,30000]": [20000, 30000]}}

dt_bucketer = DecisionTreeBucketer(variables=["LIMIT_BAL"], specials=specials)
dt_bucketer.fit(X, y)

dt_bucketer.fit_transform(X, y).head()

	EDUCATION	MARRIAGE	LIMIT_BAL	BILL_AMT1
0	1	2	9	201800.0
1	2	2	1	80610.0
2	1	2	9	499452.0
3	1	1	3	450.0
4	2	1	9	56107.0

.summary()

This gives the user a simple table of the columns and number of (pre)buckets generated by the bucketer. The information value and dtypes are also given

dt_bucketer.summary()

	column	num_prebuckets	num_buckets	IV_score	dtype
0	EDUCATION	not_prebucketed	not_bucketed	0.057606	int64
1	MARRIAGE	not_prebucketed	not_bucketed	0.016267	int64
2	LIMIT_BAL	not_prebucketed	13	0.178036	float64
3	BILL_AMT1	not_prebucketed	not_bucketed	2.915613	float64

.bucket_table()

To look at the buckets in a more granular level, the bucket_table() method outputs, among others, a table containing the counts in each bin, the percentages, and the event rate.

dt_bucketer.bucket_table("LIMIT_BAL")

	bucket	label	Count	Count (%)	Non-event	Event	Event Rate	WoE	IV
0	-4	Special: in [20000,30000]	723.0	12.05	453.0	270.0	0.373444	-0.724	-0.075
1	-3	Special: =50000	676.0	11.27	518.0	158.0	0.233728	-0.054	-0.000
2	-1	Missing	0.0	0.00	0.0	0.0	NaN	0.000	0.000
3	0	[-inf, 75000.0)	462.0	7.70	313.0	149.0	0.322511	-0.499	-0.022
4	1	[75000.0, 85000.0)	319.0	5.32	243.0	76.0	0.238245	-0.079	-0.000
5	2	[85000.0, 105000.0)	330.0	5.50	241.0	89.0	0.269697	-0.245	-0.004
6	3	[105000.0, 145000.0)	566.0	9.43	436.0	130.0	0.229682	-0.031	-0.000
7	4	[145000.0, 175000.0)	449.0	7.48	380.0	69.0	0.153675	0.464	-0.014
8	5	[175000.0, 225000.0)	769.0	12.82	630.0	139.0	0.180754	0.270	-0.009
9	6	[225000.0, 275000.0)	501.0	8.35	419.0	82.0	0.163673	0.390	-0.011
10	7	[275000.0, 325000.0)	379.0	6.32	326.0	53.0	0.139842	0.575	-0.018
11	8	[325000.0, 385000.0)	350.0	5.83	287.0	63.0	0.180000	0.275	-0.004
12	9	[385000.0, inf)	476.0	7.93	409.0	67.0	0.140756	0.567	-0.022

.plot_bucket()

We have already seen that we can plot the above bucket table for a better visualisation of the buckets

dt_bucketer.plot_bucket(
    "LIMIT_BAL", format="png", scale=2, width=1050, height=525
)  # remove format argument for an interactive plotly plot.)

.save_yml()

We can save the generated bucket to a yaml file. This yaml file can later be used to generate a bucketer as we show in the create_bucketer_from_file tutorial

dt_bucketer.save_yml("my_output.yml")

Bucket mapping

If you're interested into digging into the internals of the buckets, you can access the fitted attribute features_bucket_mapping_. For example:

```python
bucketer.features_bucket_mapping_.get('pet_ownership').labels
# {0: 'cat lover, rabbit',
# 1: 'no pets',
# 2: 'dog lover',
# 3: 'gold fish',
# 4: 'other',
# 5: 'Missing'}
```